Nuclear and Emerging Technologies for Space, American Nuclear Society Topical Meeting Richland, WA, February 25 – February 28, 2019, available online at http://anstd.ans.org/

Small Modular Fission Reactors for Space Applications

Enabling an Affordable, Commercially Developed Power Architecture for the Moon and Beyond

J. Stephen Herringa, Stephen Mackwellb, Christopher Pestakc and Karin Hilserd Universities Space Research Association (USRA) 7178 Columbia Gateway Drive Columbia, MD 21046

a 208-526-9497, [email protected] b 281-795-2856, [email protected] c 216-650-3842, [email protected] d 410-740-6232, [email protected]

Introduction advantageous, offering NASA and industry the The National Aeronautics and Space Administration opportunity to share both the risk and rewards. (NASA), other international space agencies, industry, and private entrepreneurs have plans to Common to all these robotic and human activities is robotically explore planetary bodies and to return the need for abundant and reliable electrical power. humans to the Moon within the next decade, More capable robotic exploration, ISRU, and followed by human visits to Mars. Performing these sustained human presence will require electrical activities via public/private partnerships is highly power in the 40 kW to 150 kW range that is continuously available throughout the entire day/night cycles of the planetary body being Planned Future Human and Robotic Mission explored. In the case of the Moon, a power plant Activities capable of meeting this need would form the basis • Characterization of planetary surface for establishing commercial electrical utility services morphology, geology, and on the lunar surface. Such services will jump-start geochemistry, including detailed the exploration, resource mapping, commercial exploitation, and by a mapping of mineral deposits and in situ broad mix of public and private users that include resources; space agencies, industries, adventurers, and • Exploration of surface and subsurface entrepreneurs. locations, including craters, lava tubes, and perhaps subsurface oceans; To address the challenges and opportunities of • Collection, return, and analysis of establishing in-space commercial electrical utilities, samples from the lunar, asteroidal and Universities Space Research Association (USRA) Martian surfaces and interiors; recently began an internal research and development (IRAD) project to perform a concept study of a new • In situ resource utilization (ISRU) for small modular fission reactor (SMFR) targeted for construction, life support, and propellant use on the Moon. SMFRs have significant production; advantages over other potential power sources being • Prospecting and mining of lunar and considered for the Moon. asteroidal resources for utilization in space and/or possible return to Earth; While photovoltaic systems will play an important and role in future space power architectures, their • Construction of outposts, settlements, inability to generate power during the long lunar habitats, and other infrastructure. night (which can last for up to 14 Earth days) or in shadowed regions places significant limitations for most applications envisioned on the Moon. Photovoltaic systems become impractical when accounting for the mass of the energy storage components (batteries, flywheels, etc.) required to provide electrical power through the long lunar night. On Mars, photovoltaic systems are also limited by solar intensity and atmospheric dust.

Similarly, radioisotope thermoelectric generators (RTGs), like the one shown in Figure 1, have significant limitations for planetary and deep space robotic and human missions. RTGs, which operate using exotic Pu-238 fuel, typically provide only a few hundred watts of electric power. The production of Pu-238 has only recently restarted in the United States in response to the need for deep space power sources for robotic missions. However, production is very expensive and slow, so missions requiring multiple RTGs are generally not feasible. Thus, while RTGs are suitable for the modest power requirements of many robotic missions, they are inadequate for the requirements of more demanding robotic missions and sustained human exploration. Figure 1: The Multi-Mission Radioisotope Thermoelectric Generator (MMRTG) for the Curiosity Rover Mission to

USRA’s SMFR IRAD Project Design Mars prior to assembly into the rover. Source: NASA

Characteristics

• Uses NASA’s Kilopower design as the State of Technology Readiness for Small Modular starting point; Fission Reactors • Delivers 40-150kWe on the lunar Large, high power fission reactors utilize mature surface; technology that is currently in use throughout the world in power and naval propulsion applications. • Operates for a minimum of 20 years However, because of their size, these existing unattended; and reactors are not feasible for use in space. There is a • Uses Low (LEU: concerted effort on the part of Department of Energy <20% U-235) U-Mo fuel. (DOE) to develop SMFRs for terrestrial applications, with an output of 300 MWe or less. The general motivations for the development of Using a variety of possible energy conversion SMFRs have been to address the evolution of market techniques, SMFRs can provide both the heat and demand for electric power with smaller, modular electricity needed for human and robotic activities, additions rather than introducing a massive step in including outposts, habitats, fleets of exploration capacity of 1000 to 1600 MWe. Smaller units also rovers, ISRU systems, and excavation/mining are more amenable to factory fabrication of operations. The study will also examine what aspects standardized modules such that the engineering and of the SMFR design are applicable for use on Mars tooling costs are spread over a production run of and other deep space missions. dozens (or perhaps hundreds) of plants.

This white paper describes the concept study that NASA is currently performing research and USRA is performing that could lead to development development on SMFR technologies for space of an SMFR for use on the Moon by 2028, or applications as part of the Kilopower project. sooner, if properly resourced. Kilopower is focused on showing the proof-of- concept of a fission reactor using highly enriched

2 uranium (HEU) (>20% U-235) with an initial output nearby humans from cosmic and reactor radiation. of 1 kWe and the potential to scale up to 10 kWe. The clever multiple use of materials moderator, neutron reflector and radiation shield could reduce USRA’s SMFR IRAD project is to study the the mass penalty in using LEU. The radiating requirements to build upon Kilopower’s successful 1 surfaces needed by the heat sink might also act as a kWe proof-of-concept experiment to build a cost- neutron reflector. effective, commercially developed SMFR that could deliver the electric power levels needed (40k- Initial Results of the SMFR Study 150kWe) on the lunar surface to perform extensive USRA asserts that the mass penalty associated with exploration, resource prospecting, and ISRU an LEU-based SMFR for the lunar surface system, activities. compared to an HEU-based system, is acceptable when the total cost, logistical complexity, and Use of Low Enriched Uranium security implications of HEU are fully taken in A significant difference between USRA’s SMFR account. To verify this assertion, USRA’s initial step IRAD project and NASA’s Kilopower project is that in the SMFR study was to develop mass estimates USRA is focused on developing an SMFR that uses for LEU and HEU systems of equal power output. low enriched uranium (LEU, < 20% U-235) as its heat source. While HEU has several advantages over A comparison of the masses of two sizes (1kWe and LEU (it requires a smaller mass of , 10kWe) of fission reactors for space applications allows the use of a wider range of materials, and was performed in 2017 by Poston and McClure1 of reduces the need for a moderator), there are Los Alamos National Laboratory. Using this report significant problems associated with HEU that are as a baseline point of reference, USRA analyzed the insurmountable from a commercial perspective. seven main components of the reactor and scaled them to the higher power levels that will be required HEU has sufficient U-235 to be usable for nuclear to power significant exploration, prospecting, ISRU, weapons. The manufacturing, security, and safety and habitation activities on the lunar surface. costs associated with the production, transportation, storage, and use of HEU make it commercially The various components of a surface-power reactor impractical. HEU by its very nature is the property scale quite differently with respect to the output of the U.S. federal government and must remain in power. For instance, once a critical configuration is the government’s control throughout the entire life achieved, the fuel mass of the reactor increases cycle of the fuel. Use of HEU essentially eliminates slowly with respect to power, primarily to the possibility of commercially owned and operated compensate for the insertion of additional heat pipes reactors, either in space or on Earth. into the core. On the other hand, the mass of the heat rejection system scales slightly less than linearly Conversely, LEU is much more difficult to use for with power, since the radiator area required scales nuclear weapons and has significantly fewer linearly with power, while the piping and structures restrictions associated with its use. The cost and needed increase less rapidly. availability of LEU make it a good candidate fuel for SMFRs. The reactor design challenges associated In order to assess the trends shown in the LANL with using LEU fuel will be thoroughly addressed in White Paper, USRA compared the two conceptual the concept study. reactor designs at 1 kWe and 10 kWe using Uranium 7 wt% molybdenum high enrichment uranium fuel The use of LEU fuel will require a larger mass of and the two designs using the same fuel alloy fuel for criticality. A tactic to reduce the mass containing low enrichment (i.e. 19.75%) uranium. disadvantage of LEU is to analyze the isotopes resulting in parasitic neutron absorption and to Making the assumption that the masses of individual evaluate the feasibility of removing those isotopes. components can be scaled with output power by the equation, There will probably be other structures near the reactor, either for storage or as a shield, to protect

3 ! !! ! (1) �!! = �! Using the scaling exponents shown in Table 1, the !! ! component masses, m , at another output power, P , where P1 and P2 are 1 kWe and 10 kWe and mi1 and ik k can be estimated using equation 3, mi2 are the masses of the ith component (e.g. Heat

Rejection, Gamma Shield, etc) at P1 and P2. The !! masses for each of the components are shown in !! (3) �!! = �! Figures 1 and 2 of Poston and McClure (Ref. 1) are !! ! shown in Table 2 highlighted in green. Rewriting equation (1), the scaling exponents, yi, are therefore as shown in Table 2. !! !" ! ! !! Extrapolation by an order of magnitude beyond the �! = ! (2) !" ! 10 kWe design considered by Poston and McClure !! should be viewed with caution, though the trends in The scaling exponents, y , for each of the i each of the components seem logical. Nevertheless, components are tabulated in Table 1 using the comparison indicates that the relative mass equation (2). advantage of HEU is significantly reduced at the Table 1. Scaling Exponents 40kWe to 150 kWe range that is needed for large

bases on the Moon or for ISRU. Scaling Exponents for various components, based on 1 kWe and 10 kWe designs Four points can be drawn from these comparisons: yi i Component HEU LEU 1. At 40 kWe the mass difference is 1 Fuel 0.06 0.06 approximately 33%, and at 150 kWe the 2 BeO 0.30 0.03 difference is down to just 15%. 3 Reactor Components 0.38 0.11 2. The mass difference remains at about 900 kg 4 Neutron Shield 0.30 0.18 throughout the range of powers. 3. The mass of the fuel itself is nearly constant 5 Gamma Shield 0.64 0.56 from 1 kWe to 150 kWe. 6 Power Conversion 0.60 0.64 4. About half of the numerical mass difference 7 Heat Rejection 0.90 0.90 is in the fuel itself and the remainder is in the gamma and neutron shielding. The scaling exponents appear to be logical in that core mass is nearly independent of output power: These estimates based on Poston and McClure’s once a critical configuration has been achieved, the calculations show that HEU has its best advantage at output power is dependent on heat removal low powers, as Poston and McClure qualitatively capability rather than the amount of . state in their conclusion. Given the industrial, On the other hand, for a given heat sink temperature, political and security incentives for using LEU and the radiator area needed is directly proportional to the decreasing relative mass difference at output the output power. powers that are required for lunar and Martian bases Table 2: Scaling of SMFR components for specific output power levels of the system.

4 as well as for ISRU, USRA believes that further reduces neutron absorption. This portion of the study research on a LEU-based SMFR is definitely will focus on material selections that best meet warranted. SMFR design objectives.

Next Steps in the SMFR Study Thermal Control Approach: Given the number of parameters in the optimization In the overall systems engineering of the reactor of an SMFR for use on the lunar surface and the power system, the need for effective heat sinks is desire to minimize challenges to implementation of also a challenge. While terrestrial power plants can space reactor capabilities on planetary missions, a use a convenient lake or river as a heat sink, orbiting more thorough study is needed to evaluate design reactors have to rely on blackbody radiation to the choices that will lead to a viable strategy within the 4 K background of space. As a rough comparison, a coming decade. This more detailed study, scheduled surface at 300 K radiates 460 W/m2 to space, while a for completion in June 2019, will focus on three vertical surface in still water at 300 K dissipates areas of the SMFR design. 14 kW/m2 - a factor of 30 greater. A blackbody radiator would have to be at 700 K to dissipate heat One of the important components of the study will at 14 kW/m2. Thus, the careful design of the heat be the modeling of the depletion of the U-235 sink for any orbital electrical generating system is content and the in-situ breeding of Pu-239 over the critical. long, relatively low power, operating life. Table 3 shows the consumption of U-235 after 20 years for The Kilopower reactor, which uses HEU, is cooled various power levels if no breeding is included. using liquid sodium heat pipes clamped to the outside of the solid metal core. Larger reactors Table 3. Uranium Consumption would require heat pipes in the interior of the core in Uranium Consumption in Surface Power Reactors order to reduce temperature peaking. Thermal to Electric 25% Conversion Efficiency Planned mission 20 years The Jupiter Icy Moons Orbiter (JIMO) program proposed a 200 kWe gas-cooled Brayton cycle Power kg decrease in enrichment, percentage pts (no breeding) Burnup (MWth-d/kg) (kWe) fisssioned reactor for the 20-year mission. The Naval Reactors HEU LEU HEU LEU branch of DOE conducted the design study from 1 0.031 0.089% 0.009% 0.835 0.083 5 0.156 0.406% 0.040% 3.802 0.377 2003 to 2005. The JIMO reactor design will serve as 10 0.312 0.779% 0.077% 7.305 0.721 a valuable reference point in the trade study. 20 0.624 1.498% 0.147% 14.034 1.381 40 1.247 2.877% 0.282% 26.963 2.643 60 1.871 4.215% 0.412% 39.504 3.864 A reactor on the lunar (or Martian) surface might use 80 2.494 5.527% 0.540% 51.801 5.059 subsurface deposits of ice in high latitude craters as 100 3.118 6.820% 0.665% 63.919 6.235 a heat sink. The walls of the craters shade the 150 4.677 9.993% 0.973% 93.650 9.115

Selection of Fuel Cladding and Structural Specific Design Choices for Study by Materials: USRA’s SMFR Team The choice of fuel cladding and structural materials • Selection of fuel cladding and might also be made to achieve better neutron structural materials; economy, especially if certain isotopes are removed. For instance, in the design of an LEU nuclear • Thermal control approach; and thermal propulsion reactor, Patel et al. (2016)2 found • Reactor configuration options. that the removal of W-186 in the W-uranium dioxide (UO ) fuel was necessary The use of N-15 (0.4% of 2 interior of the craters from direct solar exposure. natural nitrogen) in high-density, high-temperature Movable shades over the radiating surfaces might uranium mononitride (UN) fuel both avoids (n, p) allow the heat sinks to be more effective during reactions by N-14 and greatly reduces the production times of solar exposure. of radioactive C-14. Likewise, the removal of Mo-95 from high-temperature structural reactor components

5 Reactor Configuration Options: maybe farther, become seriously power-limited in The Kilopower design is compact, with the core instrumental capability and survivability in the outer itself about the size of a roll of paper towels and the solar system and in regions with extended periods of heat removed to the energy conversion system via no or little solar radiation, like the two-week night heat pipes containing counter-flowing sodium liquid on the Moon or the dark side of Mercury. and vapor and a central wick. This study would consider similar configurations using an LEU core Space nuclear reactors for power and propulsion with a surrounding beryllium reflector. The have been studied for three decades, with significant beryllium reflector returns neutrons back toward the progress in developing the capability. Various core with a minimum loss due to absorption. prototypes have been developed but then abandoned because of lack of continued funding (e.g., the SP- Since the LEU core is >80% U-238, it is important 100 space reactor power system). The most recent that absorption by the U-238 resonance at 6.7 eV be advancement in space reactor development has been minimized. One technique is to use thick sections of the Kilopower project, which has seen significant metal fuel and little moderation such that the neutron development in the past few years. flux near 6.7 eV was depleted at the surface of the fuel and the flux in the interior of the sphere was not In our IRAD project, USRA is addressing the design further depleted by the 6.7 eV resonance since issues for an SMFR for use on the lunar surface that uranium is a poor moderator. Thus, the annular can be developed commercially, thus enabling configuration of the Kilopower core could be greater industry participation in current and future retained, though the LEU core would have a initiatives for the exploration, habitation, and significantly higher mass than the HEU design. exploitation of the Moon. Our study will build upon the progress made by NASA’s Kilopower project Higher power versions would require more heat and make use of the DOE’s experience in the pipes, predominately on the periphery with only a development of compact reactor systems. few in the interior of the core. Component dimensions and the moderator material of choice will be the major configuration options in this study. REFERENCES Though beryllium is the probable material of choice for the reflector, graphite will be considered. 1 Poston, D. I. and McClure, P. R. “White Paper – Use of LEU for a Space Reactor,” LA-UR-17- 27226, 2017-08-11 2 Study Deliverables Patel, V., et al. (2016) Comparing Low Enriched (Scheduled for Completion by June 2019) Fuel to Highly Enriched Fuel for use in Nuclear Thermal Propulsion Systems, in 52nd Design concept(s) for one or more • AIAA/SAE/ASEE Joint Propulsion Conference, configurations; American Institute of Aeronautics and Astronautics. • Mass and volume estimates; and • Cost estimates.

Conclusion Exploration of the solar system, whether by robots or humans, is highly constrained by the need for adequate sustained power-generation systems to operate instruments, maintain appropriate thermal environments for equipment or humans, and enable resource extraction and processing to sustain the mission. Even robotic missions, which can generally operate using solar electric power out to Jupiter and

6